stromboli documentation
stromboli is an open-source AI SRE agent you host yourself. When an alert fires, it gathers context from your logs, metrics and deploys, proposes a mitigation within a few tool calls, then looks for the root cause and posts both to Slack.
These docs follow the Diátaxis framework. Each section answers a different need, so pick the one that matches what you are trying to do:
| Section | Read it when you want to… | |
|---|---|---|
| 🎓 | Tutorials | learn stromboli by doing, step by step, with a result at the end |
| 🛠️ | How-to guides | get a specific job done: connect Axiom, set up Slack, turn on machine actions… |
| 📖 | Reference | look something up: a CLI flag, an environment variable, a catalog field, an API message |
| 💡 | Explanation | understand how stromboli works and why it is built this way |
Start here
New to stromboli? Do the tutorials in order:
- Your first investigation, locally: run stromboli with a demo app that breaks after three minutes, and watch it investigate.
- Connect a real service: deploy stromboli, connect your own logs, describe a service and send it an alert.
Map
Tutorials
How-to guides
- Install and run: self-host in production, deploy on Railway, deploy on Fly.io
- Connect data: logs, metrics, deploys from GitHub, Vercel, ship Fly.io logs
- Describe your system: manage the catalog in git
- Receive alerts and report: send alerts to stromboli, set up Slack
- Tune: choose models, enable health checks
- Act: enable machine actions, approve actions from Slack
- Integrate: use stromboli from an MCP client, script the CLI
- Troubleshoot
- Develop stromboli
Reference
- CLI
- Configuration (environment variables)
- Service catalog
- Connectors
- Metric presets
- Alert webhooks
- HTTP and ConnectRPC API
- Agent tools, mitigations and reports
- Machine actions
- MCP server tools
- Database schema